An analysis of predictors of success in part II MFPHM

Abstract

Background: The examination for Part II of the Membership of the Faculty of Public Health Medicine (MFPHM) has been held since 1974. Candidates and examiners have hypothesized a number of factors as likely to increase the chances of success in the exam. This study formally investigates factors which may predict a candidate's success at their first attempt. Methods: Routinely available data on candidates sitting the exam from March 1996 to June 1999 were examined initially by univariate analysis. A logistic regression model was then constructed entering all variables identified as statistically significant as well as the following variables initially thought to be relevant (a priori hypotheses): age, sex, university versus NHS training location; splitting competencies evenly across two reports; and training within North Thames deanery. Results: Four factors were identified as statistically significant within the logistic regression model. These were: possession of either MRCP (odds ratio (OR) of failure = 0.33, p = 0.003) or MRCGP (OR = 0.45, p = 0.016), success at UK Part I at first attempt (OR = 0.42, p = 0.014), and North Thames training (OR = 0.44, p = 0.048). Conclusion: Possession of MRCP or MRCGP, or passing UK Part I MFPHM are factors not normally open to change by candidates once they begin working towards Part II MFPHM. Training in North Thames was also identified as an independent predictor of success. It would seem advisable to investigate in further detail whether candidate selection (not adjusted for in the model) or aspects of the North Thames training programme are responsible for this observed phenomenon.